Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!think!harvard!seismo!munnari!basser!john
From: john@basser.oz (John Mackin)
Newsgroups: net.sources.d
Subject: Re: YASS (yet another shell script)
Message-ID: <553@basser.oz>
Date: Sun, 23-Mar-86 08:53:16 EST
Article-I.D.: basser.553
Posted: Sun Mar 23 08:53:16 1986
Date-Received: Tue, 25-Mar-86 05:23:12 EST
References: <620@ssc-vax.UUCP>
Reply-To: john@basser.oz (John Mackin)
Distribution: net
Organization: Dept of Comp Sci, Uni of Sydney, Australia
Lines: 34
Summary: not 'head'!

In article <620@ssc-vax.UUCP> in net.sources, aims@ssc-vax.UUCP
	(John Daley) writes:

> This is a 'simple' little shell script I whipped up, out of need, to
> print a specific line of a specific file.  Yes, I chose to use the B
> shell, but it gets the job done, right?.

[essentially:]

>        head -$1 $2 | tail -1

> Anybody out there got a better way to do what my script does?

Yes.

Not to put too fine a point on it, and not meaning to be insulting,
but this is terrible.  First of all, it's intensely non-portable.
Many systems do not have ``head'', which is a Berkeley-ism.
Furthermore, imagine something like "line 4160 /usr/lib/news/log".
How much data is going to go down the pipe?  A lot.  A hell of a lot.
Scandalously inefficient.

A much more efficient way to print one line of a file, given its
number (as $1, filename as $2), which should be portable to every UNIX
system, is the following:

	sed -n -e $1p -e $1q $2

And you can surround that with as much sugar as you like.

John Mackin, Basser Department of Computer Science,
	     University of Sydney, Sydney, Australia

seismo!munnari!basser.oz!john	john%basser.oz@SEISMO.CSS.GOV